CCJ 4700 Exam 1 Study Guide Errors in Casual Observation Misinterpret or misperceive events we re bad observers See patterns in non representative observations don t assume or generalize Ignore events that contradict conclusions Illogical explanations for contradictory events Bias because of influence of politics ideologies beliefs etc Plural of anecdote is NOT data Foundations of Social Science Theory systematic explanation of a pattern that answers why Too abstract to test EX Crime is a product of weakened bonds to conventional society Hypothesis testable statement of an expectation about the world What we test Can be measured and later refined through data EX Children who dislike school will be more delinquent Terminology Aggregate group of multiple units Group with things in common EX Group of Caucasian Variable Any entity that cane take different values EX Race age and in agreement with statement I enjoy research methods Independent variable causes outcome of interest Dependent variable outcome depends on independent variable Attribute characteristic of a unit of observation A specific value on a variable EX Caucasian 25 Strongly agree with statement Causality 1 The cause is correlated with the effect 2 The cause precedes the effect in time 3 There is no alternative explanation for the association If not all three criteria are met in a study then there is no cause for causality Necessary cause one that must be present for the effect to occur EX Cannot be convicted without first being arrested Sufficient cause one that guarantees the effect of the interest EX If you enroll in CCJ4700 then you will spend time in Bellamy Partial cause one of multiple factors causing the outcome Probabilistic cause one that changes the likelihood of the effect upping the odds Validity Truth or accuracy of a measure or inference How confident should we be that our conclusion is correct 1 Statistical conclusion validity Are the cause and effect really correlated Is the association large enough to be trustworthy Internal validity Is the relationship due to the effect of some other factor 2 3 External validity Would we find this again under different circumstances Can it be applied in other conditions 4 Construct validity Do our variables represent the process we meant to represent How things are worded in study and what is really being measured Unit of Analysis The what or who being studied Two Fallacies 1 Ecological fallacy drawing conclusions about individuals from data about groups 2 Individualistic fallacy drawing conclusions about groups from data on individuals Time Ordering and Causality changing standards for evidence 1 Cross sectional data collected at a single time point one shot data 2 Logical inferences could the effect reasonably come first 3 Retrospective accounts of the past ask respondent to report on the past 4 Collect information about multiple time points Prospective follows cases forward through time Longitudinal uses information about multiple time periods broader 1 Trend study observing a phenomenon in an entire population at multiple time points How concepts change over time institutions not individual people 2 Cohort study observing a group of units who experienced a transition at the same time at multiple time points How generations change at transitions such birth year graduation etc 3 Panel study observing the same units at multiple time points How specific people change over time Measurement Concept an abstract idea expressed as a word Some concepts are east to define homicide while others are not prejudice We link theory and research by defining concepts Operational definition specific activities or measures used to observe a concept Validity is this a true accurate measure Face measured in other s judgment if it looks like it Criterion agrees with an accepted measure of the same concept Construct relates sensibly to measures of different concepts Reliability is this dependable will it give the same answer when tested again Test retest when applied again Inter rater when applied to different people Internal consistency when multiple items measuring the same thing agree with each other Good measures are valid and reliable but sometimes if measures yield the same result it doesn t always mean its correct Measuring crime has variables whose attributes must be Exhaustive each case can be classified in terms of one of the attributes Ex When taking a test and none of the answers applies adding Other as an option fixes this problem Mutually exclusive each case can be classified in terms of one and only one attribute Ex When taking a test and two answers seem correct adding More than One fixes this problem Levels of measurement describes a quality of a measure s attributes Nominal attributes are names or labels Ordinal attributes can be rank ordered Interval equivalent distance between consecutive attributes Ratio has a true zero value Crime Reports The best data source depends on the information you need Uniform Crime Reports UCR FBI tallies crimes reported to the police Two Parts 1 Part 1 offenses indicates level of crime in the US serious regularly occurring crimes reported to police 2 Part 2 less serious crimes including those suspects arrested charged Supplementary Homicide Reports SHR Police reported details about individual homicides National Incident Based Reporting System NIBRS Converts UCR to an incident based measure More details and more offense types National Crime Victimization Survey NCVS Longitudinal panel study of households Respondents answer screening questions Suggests that UCR misses half of violent crimes and 2 3 property crimes NOT a good source of information on homicide but the best source of information on sexual assault Self Report surveys The individual is the unit of analysis in this survey Respondents are asked to report their own criminal behaviors Reply on anonymous and confidentiality Can collect other information about the offender Self report crime rates are much higher than UCR or NCVS crime rates Quiz 1 I proposed that unattainable economic goals create pressure to offend I surveyed FSU seniors asking them how certain they were that they would land a well paying job after graduation and how many times in the past year they had stolen something worth more than 50 I found that respondents who felt less certain that they would land high paying jobs reported stealing more frequently I concluded that unattainable economic goals lead people to
View Full Document
Unlocking...